What Clients Actually Pay For

Most of the sites I have shipped this year were for service businesses: emergency plumbers, electricians, a flood hazard map for Davao. Different trades, one shared reality — the client does not care what is under the hood.

The brief is never the brief

A brief arrives as "we need a new website." What it means, every time, is one of three things:

• The current site does not show up when someone searches at 2am with a burst pipe.
• The form goes to an inbox no one opens.
• It looks like it was built in 2011, and that costs trust before anyone reads a word.

None of those are design problems, and none of them are solved by a rewrite alone. Diagnosing which one you are actually being hired for is most of the job.

What ends up mattering

Speed on a bad phone

The customer is standing in a flooded kitchen on a two-bar connection. That is the real test environment. Static rendering, compressed images, and no client-side JavaScript in the critical path do more for conversion than any animation.

The path to contact

Every page needs the call button in thumb reach. On the emergency trades sites the pattern that worked was blunt: a sticky call bar on mobile, phone number in the header, and a form that asks for three fields and nothing else.

A contact form with nine fields is a filter. Sometimes you want a filter. For emergency callouts, you never do.

Being findable

Structured data, a real sitemap, honest metadata, one page per service and per area served. Unglamorous, and the single highest-leverage work on the whole project.

How I set them up now

1. Next.js, statically rendered. Service pages change monthly at most. There is no reason to compute them per request.
2. Content in files, not a CMS. For a twelve-page site, a CMS is a subscription plus a login the client will forget. Files in the repo are faster to edit and impossible to lose.
3. One deploy target, one dashboard. Fewer moving parts means fewer 1am messages about something being down.

The part I got wrong

For a long time I optimized the wrong axis. I would spend a day on a page transition and ten minutes on the metadata. The transition is what I saw; the metadata is what the client's customers saw.

The tell is simple: if you cannot explain a piece of work in terms of a call the client received, it was probably for you, not for them. That is allowed sometimes — just be honest about which one you are doing.